Performance of the NSSL Hail Detection Algorithm for Multicell Storms Over the Coastal Southern Plains

摘要

A study of the National Severe Storm Laboratory's (NSSL's) Hail Detection Algorithm (HDA) has been performed for over 20 different multicell storm systems that occurred over the coastal region of the southern Great Plains. Since the HDA was developed using a data set mostly from the Central Plains where many of the thunderstorms are supercellular, this study addresses the regional and storm morphological influences on the algorithm performance. To enable use of ground validation from spotter networks, only the severe hail category was studied. While the NSSL HDA performed quite well in most instances, there was a tendency for the algorithm to over-warn for severe hail. Certain adjustable parameters within the HDA were tested for accuracy and optimized for the League City National Weather Service Office in Houston, Texas. These parameters include the Warning Threshold Selection Model (WTSM), the optimum population density threshold, and the Probability of Severe Hail (POSH) function.